I think we're in agreement that sleep is an issue for a lot of kids, but the devil is in a ton of details about how society is organized.
To comment on your points...
> 1) They wouldn't need to get out any later. Most of the school day is basically baby sitting and they could easily fit the actual teaching in <4 hours.
I agree that they wouldn't need to get out any later and that most of the school day is a waste and can be radically compressed, but I sincerely doubt that this is what would happen. If we're talking about showing up to school later and leaving at the same time, I'd say that's a big win over the current situation. Please tell me I've misread things and this is what we're talking about. I'd be happy to have that be the outcome, but I doubt it.
> 2) There have been bunches of studies that show teenagers don't get proper sleep specifically when they get up early. There is something about that age that requires sleeping later to get good rest. No matter when they go to bed.
I've seen similar tendencies in the studies I've come across as well, but I'm not convinced that the stats are interpreted correctly based on human nature. My thinking is that the people who need sleep the most probably aren't going to go to bed at a sane time and end up with a large net sleep gain. That said, I'm a little open-minded about this point applying to many (but not all people)
> 3) Who says their waking coincides with the current times? Doesn't for us. Something like 9:30 or 10:00 would.
Everybody's job and life situation is different, but I'd bet that a lot of lower income people have to deal with getting to work earlier, and have more transportation issues (trying to connect to multiple busses and trains, etc) and more complex childcare requirements necessitating less overall time flexibility in the morning.
I'd wager it's largely the technical, white-collar class that can show up later and generally has a quicker path to work (work in pajamas in living room, or hop into the car in the driveway and drive to work with no connecting delays)
